I worked on a BI reporting product on the use of Power BI/Fabric platform resources and the resulting costs generated by individual business units.
In addition, we provided formal recommendations and tools to support effective platform management, and introduced new capabilities for the Fabric platform.

The benefits of the project are:

• Tools providing information on resources used and costs generated
• New Fabric capabilities enablement
• Platform management support

I worked on migration of a data product for international leader in its industry. The product is a model for tracking user adoption and engagement with Power BI initiatives and products across the organization. On further stages the model has been enriched with data related to the use of Power BI resources.


The benefits of the project are:

• Increased performance:
  • Power BI reports responsiveness with Direct Lake
  • Processing time reduced from 8 h to 1 h
  • Data availability improved from T–2 days to T–1 day
• Improved error and fault tolerance
• Significantly increased data reliability
• Model enriched with detailed insights about resources used and costs

The solution is fully based on Microsoft Fabric platform, using data pipelines, Python notebooks, data warehouse and direct lake semantic model with Power BI reports.

The project involved implementing a recommendation algorithm for order splitting (knapsack problem) to optimally divide the order based on product parameters and available transport fleet. The algorithm is used on the wholesale platform of a leading FMCG company worldwide.


The benefits of the project are:

• Automatic recommendations for freight forwarders that replaced their manual work
• Optimized transportation resources, reducing shipping costs

The solution is based on knapsack problem algorithm implemented in Python and deployed using Azure Machine Learning.

Preparing a template for an application based on GenAI services, which, based on a chat or telephone conversation, prepares a transcription, a summary, verifies and evaluates the content of the conversation based on guidelines and proposes further actions.

The solution is based on Python with microservices architecture and Azure Cognitive Services.

I started the project as a Data QA Specialist, but from the very beginning I was fascinated by the role of a Data Engineer. In addition to QA tasks, I built data pipelines and helped with data modeling, eventually becoming the lead Data Engineer.
The product was a data warehouse built in a medallion architecture, designed to enable efficient processing of sales data from sources with limited availability while maintaining high scalability.


The benefits of the project are:

• Unified data warehouse with aggregated data related to the company's core business. Power BI reports that analyze sales of key products.
• Instant analysis of sales data that allows to respond to key products and customers.
• Enabling the building of custom reports based on the data model.

The solution was based on Azure Data Factory and Synapse Analytics services, and serverless computing.

It was my first professional commercial project and first experience with cloud computing. Worked as Data QA, responsible for validating data in the model, ETL processes, and automated validation rules within daily processing.


The benefits of the project are:

• Centralized data warehouse for Standard Audit File for Tax (SAF-T) reporting
• Reduced hundreds of hours of manual work by people to a few hours of result verification
• Minimized the number of errors that required subsequent correction

The technology stack was Azure Cloud: Data Factory, Synapse Analytics with Delta Lake, and SQL Server. Most of the code is Python notebooks and stored procedures.